Early lessons learnt from cash transfer interventions in post-Matthew Haiti

United Nations Development Program Haiti’s Country Office Poverty Reduction Unit Early lessons learnt from cash transfer interventions in post Matthew Haiti Version 6.0 Consultant: Yadira Díaz May 2017 2 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) Contents 1 INTRODUCTION.................................................................................................... 5 2 A REVIEW OF CASH TRANSFER INTERNATIONAL LEARNT LESSONS USEFUL FOR POST MATTHEW INTERVENTIONS .............................................................................. 7 2.1 Documented impacts of cash transfer programs in emergency contexts: overview ......... 8 2.2 When a cash intervention is pertinent? ......................................................................................... 10 2.3 Deciding the type of cash intervention to implement ............................................................... 11 2.4 The size of the cash grant.................................................................................................................... 13 2.5 Do cash transfers encourage community-sharing systems? ................................................... 13 3 EARLY LESSONS LEARNT FROM THE UNDP’S CASH FOR WORK POST MATTHEW INTERVENTION IN HAITI........................................................................................... 14 3.1 General overview of the UNDP program ........................................................................................ 14 3.2 Targeting of cash transfer beneficiaries ........................................................................................ 17 3.3 Health and socio-demographic vulnerable populations .......................................................... 19 3.4 The gender perspective ....................................................................................................................... 21 3.5 The size of the cash for work grant ................................................................................................. 22 3.6 Interoperability with other programs ........................................................................................... 25 3.7 The role of local authorities ............................................................................................................... 26 3.8 Allocation of resources and geographical coverage of the program.................................... 26 4 OTHER TYPE OF CASH INTERVENTIONS: A MULTIPURPOSE GRANT SIZE DESIGN FOR HAITI ................................................................................................................ 27 5 REMARKS AND STEPS AHEAD ............................................................................. 31 REFERENCES ............................................................................................................ 32 ANNEXES ................................................................................................................. 33 3 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) 4 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant)  1 Introduction The third and fourth of October of 2016 the category four hurricane, Matthew, hit Haiti. According to the United Nations Office for the Coordination of Humanitarian Affairs (OCHA), Matthew affected 2.1 million of persons; population that represent in about 19% of the Haitian inhabitants. In terms of number of affected population, the most affected departments were the Sud, Grand’Anse, West and the Nippes Departments, with 775 thousand, 468 thousand, 222 thousand and 205 thousand affected persons, respectively (OCHA, 2017). The assessment of the impact of the hurricane accounts for at least 90 thousand houses destroyed or strongly damaged and 806 thousand persons left in urgent need of food or agricultural assistance. As part of the humanitarian response to Matthew, the 13th of October the United Nations Development Program (UNDP) launched a cash transfer intervention firstly in the West Department in the Grand Goâve municipality. Followed by that, during November UNDP expanded its cash transfer program to four municipalities in the Grand’Anse Department and in December to other six municipalities in the same Department. This UNDP intervention took place along many other interventions from different humanitarian partners as for example the International Organization for Migration (IOM), the World Food Program (WFP), the United Nations Children Found (UNICEF), the Food and Agriculture Organization of the United Nations (FAO), the National Directorate for Water Supply and Sanitation in the Ministry of Public Works (DINEPA), and the World Health Organization (WHO). In fact, a large number of organizations have implemented post Matthew cash activities. As seen in Figure 1, at January 2017, the number of organizations with ongoing cash activities was 22. These activities were benefiting 240 thousand persons. At the 10th of April of 2017, 53 organizations have had post Matthew activities in Gran’Anse and the Sud departments. To coordinate the activities of all these humanitarian partners, working groups have taken place in Port au Prince, Jeremie and Les Cayes. In particular, one of the working groups that has been meeting to coordinate post Matthew interventions is the cash-working group. The cash-working group consists of the different UN agencies and humanitarian partners that have been putting in place cash interventions in the Hurricane affected areas. Given the large number of organization involved in post Matthew cash activities, a cash steering committee was created to provide technical guidance to the different implementing partners and to standardize the different aspects of the interventions. This steering committee was installed in November 2016 and conformed by three cash leading UN partners (WFP, UNDP, FAO), three NGO representatives, the American Red Cross, and donors (USAID and ECHO). 5 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) Figure 1. Cash transfer post Matthew activities, January 2017 This technical report has twofold purposes, firstly to describe the main international evidence on cash transfer programing pertinent for the Haitian post Matthew emergency context; and secondly, to document the main lessons that can be learnt from the UNDP post Matthew cash transfer intervention. The lessons learnt presented in this document from the UNDP post Matthew cash transfer program were drawn from quantitative and qualitative assessments. In particular, quantitative techniques were used to critically review the design characteristics of the program. For this, the report uses the 2012 ECVMAS and the CNSA food basket prices monthly tracking. Qualitative techniques, on the other hand, were used to critically assess the processes of the program. In this regard, the evidence drew from semi-structured interviews with different stakeholders such as beneficiaries, local authorities and infield personnel was used to shed light on the early lessons that the operation of the program can provide. Worth remarking, this report is not meant to be an impact evaluation of the program, but rather an overview of the operative lessons learnt from the first six months of its implementation. It does not analyze the lessons learnt from other cash programing interventions that took place during these six months, as for instance unconditional cash transfers with nutritional purposes. The production of this report has been fully founded by the Poverty Reduction Unit at the Haiti’s Country Office of UNDP and has benefited from the discussions and comments received during the meetings of the cash steering committee and the 6 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) cash-working group that took place in Port au Prince from November 2016 until May 2017. The report is organized as follow: First, in Section two, the international background relevant for the Haitian context with regards to available lessons learnt from cash interventions in other humanitarian emergencies is summarized. Section three follows describing the UNDP post Matthew cash intervention and critically discusses the lessons learnt from it, this in light as well of the discussions that took place in the cash steering committee. The last section, Section four, presents the design of the grant size for a multipurpose grant, cash intervention that is not part of the UNDP post Matthew program but that has been discussed and analyzed during the cash steering committee meetings. 2 A review of cash transfer international learnt lessons useful for post Matthew interventions This Section presents a general overview of the main documented findings and lessons learnt from cash interventions in emergency contexts. It heavily draws from the Harvey & Baily (2011) good practice review. Cash transfer interventions have become during the last years one of the most preferred modalities because their proven ability to meet particular program objectives, while giving flexibility to encompass beneficiaries’ preferences. Still, not golden rules to follow exist when programing cash interventions, but there is however an important amount of lessons learnt and evaluations from previous cash interventions around the globe that have been useful for the post Matthew Haitian context. This section presents an overview of the documented impacts of cash interventions, then, briefly discusses when a cash intervention is pertinent, the type of cash intervention to implement and some guidelines to set the amount of the transfer. Given the growing number of agencies and counties implementing cash transfer interventions, to compile and share knowledge and learning programing experiences, a partnership organization, CaLP (the Cash Learning Partnership), was created by Oxfam Great Britain, Save the Children UK, the British Red Cross, the Norwegian Refugee Council and Action Against Hunger USA. The post Matthew Haitian context was made accessible to this particular partnership organization by 7 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant)  members of the cash steering committee who have interchanged information and experiences with CaLP.1 2.1 Documented impacts of cash transfer programs in emergency contexts: overview Several cash interventions and evaluations of them have produced an important amount of information that can guide the program design, planning and implementation of these interventions. Documented international impact evaluations can be found across several developing countries and sectors. In particular, cash interventions have proven been effective on Food security and nutrition, livelihoods, shelter, distribution of non-food items, access to basic services, household decision making, gender equity and empowerment. Table 1 below presents a brief overview of these main findings across implementing sectors. Table 1. Documented impacts of cash interventions Implementing sector or Intervention Documented impact outcome’s type Kenya. Cash transfers increased the dietary diversity among Refugee targeted households and allowed vendors to expand camps, 2007 their business (Dunn, 2009). Cash has been found to be more effective than food aid to increase frequency and dietary diversity (Harvey & Bailey, 2011). Food security Ethiopia, 2009 Cash transfers or vouchers are found not to be an and nutrition appropriate modality to ensure specialized food micronutrients to address severe and moderate acute malnutrition (Jaspars et al., 2007). Combinations of cash and food was found to be the Indonesia most effective modality to meet the nutritional objectives of the program (Skoufias et al., 2011) 1 Table A1 in the Annexes includes two focal points identified by one of the members of the post Matthew Haitian cash steering committee for future contact in CaLP. 8 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) Implementing sector or Intervention Documented impact outcome’s type In an emergency relief context cash is most likely to be spent on immediate consumption. However, cash Kenya, 2002 also can promote productive investment in context where the situation is less acute or the total amount paid is large enough (Frize, 2002). After a severe drought, cash transfers arrived in 2006, when the situation was less acute and was 85% Kenya, 2005 Livelihoods of the times used productive investments as animals (O’Donnell, 2007). Cash recipients were found to received more attention from traders to whom debts were owned. The cash grant was found used in debts repayments. Mongolia As such, indebtedness needs to be considered in the particular context when setting transfer values (Harvey & Bailey, 2011). Cash transfers to encourage families hosting households or persons in emergency could be seeing as an undermining factor of the community Sri Lanka, solidarity. However, in Sri Lanka, the Swiss aid Shelter 2004 tsunami agency Helveta found cash transfers in exchange of hosting families did not clash with cultural norms and did not undermined the community solidarity in times of disaster. Comparing the amount of the cash transfer spent on non-food items versus the basket typically include as Democratic part of the humanitarian response of non-food items, Non-food Republic of only 8% of the cash transfer spent on goods was items Congo, 2007 spent on the goods found in these kits; mostly was actually spent on goods such as clothing, mattresses and bicycle parts (Bailey and Walsh, 2007). Cash also has been found spent on accessing services Access to basic Zambia, 2005 such as school or education (Harvey and Morongwe, services 2006a). Household Household decision making was found impacted by decision Sri Lanka the cash transferred. 60% of the couples reported making taking the decision about the spending jointly rather than only by men as traditionally in Sri Lanka 9 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) Implementing sector or Intervention Documented impact outcome’s type (Sandstorm and Tchatchua, 2010). Cash transfer program directed to women was found to improve gender equity and empowerment. Adnam and Provision of services was made to ensure women Gender equity Nicobar were able to fully participate in the program as for and islands, 2004 instance childcare facilities and breastfeeding breaks. empowerment Indian Ocean Also, women were encouraged to challenge gender tsunami. stereotypes by taking up activities such as fishing (ActionAid, 2008). 2.2 When a cash intervention is pertinent? Despite the important proven impacts of cash interventions, cash programs are not necessarily pertinent in every context. The plausibility of a cash transfer program relies on the needs of the potential beneficiaries and the ability of the markets to absorb and to circulate the monetary injection that a cash program brings. According to Harvey & Bailey (2011): “Cash or vouchers will only be appropriate in situations where food or the other items that people need are available in the local markets, or can be supplied relatively quickly trough market mechanisms. In some situations, such as the early stages of a quick-onset disaster, there may be an absolute shortage of food or other items at local or national level, or markets may be disrupted. In these circumstances cash or vouchers will not be appropriate… Yet even during quick-onset emergencies markets may still be functioning, and it may be possible to deliver cash.” In consequence, the following two key questions to determine appropriateness and feasibility must be answered affirmatively to implement a cash transfer intervention: i) Are people able to buy what they need, at reasonable prices? ii) Can cash be delivered and spend safely? To answer these two questions Harvey & Bailey (2011) suggest pre-intervention assessments, to determine appropriateness and feasibility, on: • The population needs • The level of affectation of the markets • The security and the delivery options • The risk of corruption 10 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant)  • The gender and power relations within the household and community • The cost-effectiveness of the proposed cash intervention • The coordination and political feasibility • The skills, technical capacity and ability of the implementing agency to quickly respond to the emergency with the cash intervention • The agriculture seasonability The post-Matthew cash interventions have shown that the experience and size of implementing partners vary broadly, so, the exhaustiveness of these pre- intervention assessments might vary as well across contexts and implementing agencies. 2.3 Deciding the type of cash intervention to implement Four types of cash interventions are worth naming: Unconditional cash transfers, Conditional cash transfers, Cash for Work or Vouchers. While unconditional cash transfer give money as a direct grant to people without conditions or work requirements and people can spend the money as desired, conditional cash transfers stipulate certain objectives or activities that the beneficiaries need to meet or perform to be entitled to the grant. Cash for work programs can be understood as a type of conditional cash transfer, where the conditionality to the transfer is the work performed. Voucher programs refer to those where a voucher in paper or tokens or electronic cards are given to the beneficiaries and they can exchange them, either for cash or for specific products (as food for instance) or in specific vending places. In an humanitarian setting, as the post Matthew Haiti, conditional cash transfers are not the most appropriate response as this type of cash transfer imposes an important administrative and planning burden to the implementing agency, which might not facilitate the quick response required in the specific setting. Different case studies have shown that the type of cash transfer to implement depends on the objectives of the program, the context and the ability of the implementing agency to undertake one or other type of cash transfer. For the purposes of this report I am including below, as Table 2, the comparison of the four different types of cash interventions made by Harvey & Bailey (2011). 11 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant)  Table 2. Comparing different types of cash transfers Type of Circumstances when cash Potential advantages Potential disadvantages used transfer -Security situation adequate for mobilizing -Minimal -Recipients may spend the Uncondition cash transfers administrative burden money in consumption al cash -Project objectives do -Generally most cost- not linked with the transfer not restrict efficient than vouchers. project objectives. expenditures to specific goods and services. -Requires staff time to verify conditions have Conditional Specific needs have to be Agency can influence been met. cash met (e.g. shelter, small recipient expenditures -Recipients may have transfer businesses) to promote objectives other priorities for which they would prefer to use the cash. Vouchers Security concerns can be -Recipients do not -Can limit recipient choice mitigated by using directly handle cash -Recipients may have vouchers -Quality of goods and other priorities of which prices can be they would prefer to use monitored vouchers - Agencies can easily Requires more planning, influence recipient preparation and choice and promote sensitization practices -Traders not involved in the project may be disadvantaged -Traders may manipulate prices Cash for -Public or community -Dignified type of -Could disrupt local labor work works are required social assistance markets -Equipment, technical -Creates community -Can take away time from assistance and assets or facilitate other household activities supervision can be emergency response -May exclude those provided -Potential for skills without capacity to work, -Population has capacity transfer including labor poor to undertake the work households. - Capacity to maintain assets is created Source: Harvey & Bailey (2011) 12 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant)  2.4 The size of the cash grant The amount of the transfer in a cash-based intervention is defined as a function of the objectives of the program, the amount of available resources and the size of the need of each potential beneficiary. However, needs vary among population of different demographic population subgroup, as are gender or age-specific needs. Then, depending on the program objectives, different sets of households/potential beneficiaries might face a different set of needs to meet these objectives. Then, to set the amount to transfer by a program the key questions to be answered are: i) How much money do the households/potential beneficiary needs to meet the objective of the program; ii) Do we need to tailor different grant sizes according to different types of households/potential beneficiaries, according to their needs? iii) How many instalments will the program cover? iv) Do we need to adjust the grant size across time and geographical area? As the international literature agreed upon it, balance is required when setting the amount of a transfer. A too low amount of transfer would not allow the recipients to meet their needs, but a too high amount could conduce on concentrating the intervention in very few among the whole affected population. The amount of the transfer therefore needs to be tailored balancing this trade-off. For the specific case of Haiti, in the context of the cash steering committee, this consultancy jointly with OCHA designed a transfer value for a possible multipurpose grant. In this design the afore-presented four relevant questions to set the amount of a cash transfer value were evaluated. This proposed grant size for Haiti is discussed and described ahead on this document in Section 5. In general, the literature describes as the most cost-effective grant size such that it cover the amount of the household’s needs, without surplus or avoiding leaving households still in need. For the Haitian case, that approach is however, distant from the amount of available resources. The number of households and the size of their needs exceed in an important magnitude the availability of resources. For this reason, despite in about 80% of the rural population is in expenditure poverty conditions, targeting strategies are required to prioritize those households that area in most need. Section 3.2 and 3.7 present some recommendations from this consultancy to address this particular issue. 2.5 Do cash transfers encourage community-sharing systems? Evaluations have found that cash transfers are seen among participants more as an individual entitlement, whereas in-kind aid is more prone to be shared among the 13 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) community. An impact evaluation of a WFP cash transfer program in Zimbabwe evaluated the differential impact on the community of three types of transfers. One set of households received exclusively cash, another set food products and the third set a combination of food and cash. Cash was found to be share less widely than food, non-recipients benefited more from households that received food. Cash was found to erode community-sharing systems (Kardan et al., 2010). Similarly, an evaluation performed in Zambia found that cash was rarely shared, whereas food was commonly given to friends, relatives or other families as a form of gift or payment (Harvey and Morongwe, 2006b). This initial evidence suggests that cash transfer programs are less likely to strengthen community-sharing systems than in-kind transfer programs. Still, the international evidence also suggests that if the amount of the transfer is big enough households could also decide to found joint projects, as is the case of an IFRC project in Niger, which accounted 46% of the total beneficiary households as pooling a portion of the received cash from the program to found joint projects in wells and mosques reparation and building cereal banks (Harvey & Bailey, 2011). This report follows presenting the early lessons learnt from the UNDP’s cash for work post Matthew intervention in Haiti. 3 Early lessons learnt from the UNDP’s cash for work post Matthew intervention in Haiti This section starts by briefly describing in subsection 3.1 the objectives of the UNDP cash program, the population attended and its operative cycle. Then, the following subsections focus on critically describing specific characteristics of the program that require attention and the derived lessons learnt from the implementation during these past six months. 3.1 General overview of the UNDP program The post-Matthew UNDP cash for work program has been put in practice within a broader institutional strategy that contemplated: i) Risk prevention (before Matthew) and management; ii) Building local capacity and restoration of means of subsistence; and iii) Supporting economic recovery. Figure 2 below shows this strategic framework from which the cash for work program was conceived. Twofold purposes the program was created to attain. Firstly to immediate stabilize livelihoods via re-building useful and sustainable community assets and assisting 14 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) those populations most affected and in most need after the hurricane Matthew. Secondly, to strengthen the capacities of local authorities. In turn, the operative cycle of the program can be characterized by four main stages: i) Preparatory stage; ii) Selection of beneficiaries; iii) Execution of works; iv) Roll out of payments. The following paragraphs briefly describe these stages and the following sections discuss the main lessons learnt and recommendations derived from this consultancy from the UNDP operationalization of these procedures. Stage 1: Preparation stage. The preparatory stage comprises the dialogue with the local authorities, which are a key stakeholder in the whole process. This dialogue has helped the local authorities to identify the municipality needs in terms of the community assets required to be re-built as an effect of the Hurricane Matthew. Preparatory meetings take place during this stage to inform the different actors of the program and prepare the field for its implementation. As product of this stage UNDP signed an agreement with each of the municipalities targeted as beneficiaries, the sites of the works were targeted and a preparation meeting at the communal level takes place. Worth noting here that as part of the agreement with the municipality a 10% of the total cost of the program in each municipality is transferred to the local authorities to strength their capacity. Local authorities can make use of these resources as per their needs. Figure 2. UNDP post Matthew strategy Stage 2. Selection of beneficiaries. Local authorities were given, by UNDP personnel, a list of criteria to prioritize workers and based on these criteria local 15 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) authorities provided the list of beneficiaries (Workers). This list is endorsed by UNDP after qualitative assessments of some of the selected beneficiaries. Stage 3. Execution of works. Once the beneficiaries and the sites to intervene have been selected, the intervention takes place in slots of 10 days per worker and groups of 20 workers per one supervisor. Every five supervisors one controller was make available. Stage 4. Roll out of payments. Once the slots of work are finished. The payment procedures have taken place via the Haitian cooperative – Caisse Populaire. Section 3.6 ahead in this manuscript critically describes the modality used for the payments. This operative cycle was implemented in 20 municipalities affected by Matthew. In total 39,693 persons were hired for cash for work activities and only one person per family was allowed. Out of this total amount of beneficiaries, in average 90% of them corresponded to workers, 8% to supervisor and 2% to controllers. Figure 3 below plots the total number of beneficiaries of the program, which included workers, supervisors and controllers. Figure 3. Number of beneficiaries of the UNDP’s post Matthew Cash for Work program Note: Figures at May 2017. Total: 39,693. Source: UNDP, Poverty Reduction Unit. 16 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant)  3.2 Targeting of cash transfer beneficiaries During these past six months of intervention, the targeting of the program has been done, first at the municipality level, and then, within each municipality local authorities provided the list of beneficiaries. The selected municipalities are meant to be the most affected by Matthew. However, there was not available a list of specific objective criteria followed to determine how the degree of affectation was measures and how benefiting municipalities were prioritized over others. As such, the program took place by allocating a flat amount per municipality regardless of the size of its population or the affectation degree in most of the cases. This topic is discussed in Section 3.7 providing some recommendations for future expansions of the program or future implementations of similar cash transfer programs. Now, in terms of the selection of beneficiaries, targeting through local authorities proved during these six months to bring important inclusion and exclusion errors that could be avoided through better targeting mechanisms. Specifically, during the first week of November this consultancy did a process- evaluation of the program through an infield tracking of each of its processes. Among different actors within the program and stakeholders a semi-structured interview was conducted to assess the general perception and satisfaction of the program, the role of each actor in the operation of the program and the advantages and disadvantages of the program. One of the results of this evaluation pointed out that the targeting done by the local authorities has been mostly based on subjective criteria and proved to be highly captured by local interests. This targeting proved to overlook the bottom of the welfare distribution. According to ECVMAS 2012 figures, 83% of the rural population in Grand’Anse and 69% of the Sud rural population were living with less than 2 USD per day. This implies that even before the hurricane the number of poor population living in the most hit departments was already considerable and the size of their unmet basic needs as well. In general, emergency contexts affect the most to the poorest. Post Matthew Haiti is not the exception. This situation is most aggravated with the size of the need already existent in the country before the hit of the Hurricane. Then, targeting becomes important in this context to prioritize those households that are most in need over those than not necessarily are well-off but their need is significant lower than those most in need. 17 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant)  Figure 4. Expenditure poverty incidence in the Sud and Gran’Anse departments by area Source: ECVMAS, 2012 and author’s calculations. As such, the recommendation is to implement a targeting method based on layers of objective targeting criteria, where the first layer correspond to geographical targeting based on the risk mapping available and the diagnostic of affected areas by the Hurricane. The second layer corresponds to information gathered to prioritize households within each of the selected geographical areas. The smallest the geographical are targeted the most cost-effective the method, as the information required at the household level would be less costly to gather. In this regard, for the case of Haiti, municipalities are composed by section communal and section communal by localities. In the future, when a census data is available the recommendation is to undertake poverty mapping in use of the census data to select the smallest possible geographical areas based on poverty criteria. However, for the time being, as this information is not available for the country the recommendation of this consultancy is to use community based criteria to rank localities within the section communal/municipality and gather a small questionnaire on socio-demographic characteristics and deprivations that the household face. The number of household to target is constrained by the availability of resources in this case. A similar multilayer approach has been implemented successfully in Burkina Faso 2009 after a season of high food prices and unemployment (Outtara and Sandstrom, 2010). Following the experience of CARE international to target their cash beneficiaries in post Matthew Haiti, the household questionnaire is proposed to be collected by a geographical operative carried out by community members and 5% of the 18 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) collected information would be randomly selected to be validated. A more detailed documentation of this proposed method and the tools to carry it out are available under request and have been provided to the Poverty Reduction Unit in UNDP. Cost-efficacy is the goal of the targeting method and in an emergency context as the Post Matthew Haiti, too expensive and exhaustive methods would conduce to a less cost-effective and slower response, but no targeting method or a subjective criteria based one (as the one that is currently in place by the UNDP cash for work program) leads to inefficiencies that reduce the impact of the program in the most affected population. The proposed targeting method of this consultancy leads ranking and prioritizing households that suffer the largest number of deprivations at the same time and that are made by health and demographically vulnerable populations. One household member per targeted household would enroll the cash for work program in this situation then. However, what would happen with households made of only one working-aged person and children or households with no working aged population? At this point therefore, arises the question with regards to the treatment of non-economically active population that has still been hit by the hurricane and that is living in precarious conditions. Next section addresses this issue. 3.3 Health and socio-demographic vulnerable populations Culturally in Haiti there is an important awareness of the need to take health and demographically vulnerable population into account by social programs. However, in the context of a cash for work program visits to the operation of the program have proved that there are very little tailored approaches for these populations subgroups. In particular, this is the case of pregnant women, disabled population, elderly and children. In the case of cash for work programs, households consisting of these population subgroups are also the most likely to be poor and therefore also most likely to be hit severely by the hurricane. However, including pregnant women, disabled population, elderly or children in the heavy cash for work activities could conduce to non-desirable outcomes. As such, a complementary unconditional cash transfer modality is proposed to be implemented to cover households that are highly eligible but that can be considered by a cash for work program as ‘labor poor’ households. Labor poor households are defined here as those households with no working- aged population at home. In the Sud department, these households represent 5.7% of the population of households, grouping together in about 15 thousand persons. In the Grand’Anse department these households represent 6.8% of the population of households (See Figure 5). 19 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant)  Figure 5. Demographic composition of the population in Grand’Anse and the Sud department by household type Source: ECVMAS, 2012 and author’s calculations The outcome of the proposed targeting through a multilayer approach and the differentiation of the type of program by household type leads to the program design shown in Figure 6 below. Figure 6. Program design: multilayer-targeting approach 20 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant)  3.4 The gender perspective Not only vulnerable populations deserve to be addressed via a particular approach, also a gender perspective could be implemented when operationalizing cash transfer programs. In particular, cash for work programs can be seen somehow non-favorable to women because take time from other household activities making difficult the real access to the program, also because some of the work activities could be seen as dissonant with gender stereotypes of work. As such, UNDP has attempted to break these stereotypes and carried out the post Matthew cash for work program incorporating a gender perspective. This gender perspective has consisted on, firstly encouraging all intervened municipalities enrolling women within the list of beneficiaries via suggesting a desirable 40% of female quota of workers. Secondly, in the municipality of Abricot, UNDP carried out a ‘social experiment’ targeting 100% of the beneficiaries of the program as women. These two types of experiences have shown promising expected impacts with regards to women empowerment and breaking gender stereotypes that could be worth quantifying to share with the international community. As such Box 1 presents a summary of a proposed impact evaluation to carry out in the near future to quantify the impact of this experience in the relevant outcomes. Box 1. A proposed impact evaluation to share international On the other hand, the results: infield observations of the operationalization of the ¨ There is an on-going ‘social experiment’ in one of the intervened communes (100% of the beneficiaries are program and the process- women). evaluation carried out in November demonstrated ¨ The results of this experience would be valuable to that encouraging the share not only for Haiti but for the international enrollment of women is not community. The current available evidence in this regard is scarce and this on-going experience could enough to assure their shed light to encourage replicability of this type of participation, but also gender sensitive interventions in the developing world. breaking the access barriers is required. As ¨ The proposed impact evaluation: compare the differential effects of gender empowerment outcomes such, next rollouts of the across three types of intervention (groups): program are recommended to implement day care ¤ Type 1: 100% women beneficiaries (Abricot) strategies to assure single ¤ Type 2: Mixed group of workers encouraging headed households are able women to participate via an extra-fee on their to participate. payroll. Also, to encourage women ¤ Type 3: No gender sensitive intervention. participation, a strategy worth considering in the 21 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) future is the addition of a small amount in the women-wages to signal the desirability of their participation. 3.5 The size of the cash for work grant In general, one of the well-documented possible disadvantages of cash for work programs is their ability to disrupt labour markets. Whenever the wage paid by cash for work programs is higher than the equivalent wage in the market, a substitution effect might happen because workers could prefer to enrol the assistance program, rather than joining other job opportunities available in the market. As such, the international evidence recommends avoiding disrupting markets by setting a wage value in an amount slightly lower than the equivalent value in the local markets. In Haiti the minimum wage is updated every year and varies by sector and type of work. The two sectors where cash for work activities are carried out are segment B.1 and segment C.2, which correspond to: B.1) Buildings and public works; and C.2) Agriculture, forestry, breeding and fishing. At the moment of writing this report, the minimum wage, according to the official diary of the government published on May 2016, for these two sectors was 285HGD and 260HGD per day of 8 hours, respectively. This minimum wage corresponds to the value before social security payments. Also, the minimum wage in Haiti tends to informally vary across regions. In the Grand’Anse department, the Dan-Marie and Jeremie municipalities for instance, local authorities state in December 2016 that the value that they use as minimum wage was 275HGD per day. Currently, the UNDP cash for work is paying to their workers for the post Matthew activities 300HGD per day of 8 hours, as per agreed also with other implementing partners of cash for work. Taking into account that cash for work programs do not pay social security, this 300HGD paid amount can be considered slightly lower than the minimum wage. Taking these facts into account, a disruption in the labor local markets is not expected as a consequence of the cash for work activities. It is expected to be rare the case of a worker that would prefer to substitute other job opportunity for this cash for work. Also because the slots of work up to date have been assigned to be only 10 days per work. Now, in terms of the size of the slots, 10 days per worker means in average an assistance of 3000HGD per worker benefited. If we update the 2012 poverty lines of Haiti, with the IPC, the extreme is 1,758HGD per month per person. In the Haitian case, the value of extreme poverty line represents the value to purchase the subsistence food requirement per person. 22 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) However, cash for work is benefiting one person per household and households of different sizes and composition face a different amount of needs. This means in terms of the value of the poverty line, that a household consisting of for instance two persons needs 3,516HGD per month to cover its nutritional requirements. Considering that the average household size in Haiti is 5 persons per household, the amount received from the cash for work activities is not enough to cover even the amount of food subsistence. Still, covering the food subsistence for all poor and Matthew affected households requires an amount of resources not available. Then two recommendations arise to address this issue: First, a differentiation by household size of the amount of the grant in terms of the number of days that each worker is assigned could be implemented. In such a way, households consisting on 1-3 persons could benefit with a slot of 10 days of work, while households consisting on 4-6 could benefit from slots of 20 days of work and households of 7 and more persons could benefit from slots of 30 days of work. This type of differentiation would mean a total amount of the value that the household receives of about the extreme poverty line (See Table 3). One example of this type of approach was implemented in Malawi 2005 in a food and cash transfer program, where the value of the grant size was set as different across three types of households: those made of 1-3 persons, 4-6 persons and 7 or more persons. Despite the agency had fears incentivizing households to lie with regards to their composition to receive a larger transfer, the evaluation of the approach showed positive results and actually households understood such an approach as more equitable and expressed desire that the grant should have been adapted more precisely to each household size (Devereux et al., 2006). Interviews with the beneficiaries and the local authorities of the benefiting UNDP municipalities pointed out that households would prefer to work more than the 10 days assigned per workers. Table 3. Proposed slot size of cash for work by household size Source: author’s calculations based on ECVMAS 2012. 23 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) A different curse of action would consist on providing in addition to the cash for work an unconditional cash transfer to cover al least the nutritional requirement. However, beneficiaries and local authorities in the Grand’Anse and Sud departments have also pointed out that they perceive cash for work as a more dignified type of assistance than the unconditional cash transfers. According to the interviews it gives ‘self wordiness’. Second, efforts of partnership among different organizations to target the most deprived households and to complement the transfer given to them making sure these households are also the ones that benefit the most is due in the cooperation agenda. This issue is discussed further in the next section. 3.6 The modality of payment Different modalities to undertake the payments are possible in Haiti. These are: mobile money, cooperatives and Unit transfer. The UNDP’s cash for worked decided to the cooperative Caisse Populaire because its geographical and cost advantages. The mobile money was not selected as modality of payment because at the moment of rolling out the program, there was not available a cost efficient agreement with the providers that could allow firstly households without cellphone still access to the mobile money via complementarity of agreements with the provider. Also, there was not certainty with regards to the coverage of the service for rural and remote areas. Still, at the moment of writing this report, some of these conditions have changed. As such, a complete assessment of the cost of using different financial partners to operationalize the payments is needed. Given the accessibility and security constrains of the Haitian context, this assessment must not only includes the financial cost comparisons among providers but also the indirect economic costs that each modality of payments could bring. Still, aside from the modality of payments additional lessons can be learnt from the implemented distribution of payments. The report follows discussing them below. Through the cooperative, UNDP implemented slots of from 100 to 200 persons/payments per day. Despite the payments were done via the cooperative, massive operatives of payment proved to bring security and operative challenges that UNDP tailored by implementing days of no more than 200 payments at a time and per communal section - locality. This means that the cooperative facilitated the payment to the beneficiaries by geographical area to avoid imposing translation costs to the beneficiaries to claim their payment. Very few beneficiaries did not access the payment via the designated day per locality / communal section. These few cases were households that were living even further away. In these cases, the cooperation among the financial institution 24 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) and local authorities proved to be effective as they both worked together to locate the household and facilitating the payment. Other important constrain in the Haitian case, worth mentioning is the absence of identification for about 30% of the beneficiaries. This posed important operative constraints. In this case what it proved to be most operatively effective was advising the households that had this identification problem to claim their payment in the presence of the supervisor of the slot and or two witnesses with identification who can attest the identity of the worker. 3.7 Interoperability with other programs Given the size of the need across Haitian households, cash transfers can be complemented across different agencies and different forms of programing, as for instance, in-kind assistance, nutrition, technical advice and training. Complementarity among agencies and programs could be exploited towards meeting the needs of the most deprived households. One documented example in literature is the case of Malawi, where Concern Worldwide combined cash and food transfers to ameliorate the possibility that markets would not be ready to respond to the increase in demand via cash transfers and the possibility of create unwanted further inflation. According to Devereux et al. (2006), the combination of food and cash was found to provide the benefits of both, while overcoming the limitations of each one. Similar example can be found in Swaziland 2007 and 2008, where Save the children covered half of the food requirements through a food ratio and half trough a cash transfer. The evaluation of this approach showed that 91% of the beneficiaries (both the food only and the cash and food combination beneficiaries) preferred the combination of food and cash (Harvey and Bailey, 2011). For the case of Haiti, the interoperability of complementary programs is recommended to be strengthened at first instance by at least the targeting of those households that are in most need, ensuring the maximization of the impact on them. A first attempt in this direction was discussed at the cash steering committee during the post Matthew activities coordination. However, issues on the ability of the implementing partners to share the information collected from the beneficiaries arose and no further efforts were carried out. To overcome this problem, future targeting is recommended to be done firstly in use of similar questionnaires and criteria and also including always within the collected questionnaire the statement of the head of the household or the responding household members authorizing the use of the information for the purposes of targeting of social programs. This would enable agencies to share this information among each other for these purposes. 25 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) On the other hand, emergency cash transfers could be linked to long-term social protection programs. Whenever, social transfers are already in place they could be used to reach the potential beneficiaries for the emergency context. For example in the case of Mexico and Brazil this approach was implemented to respond to a crisis in 2007 and 2008 of high food prices, by incrementing the size of their long term cash transfer programs, Oportunidades and Bolsa familia, respectively (Harvey and Bailey, 2011). In Haiti, this could be the case of for instance, the Kore-Lavi program, which already covers particular municipalities. 3.8 The role of local authorities Along to the large number of organizations that have worked in post Matthew activities, the long-term local capacity is been built with the implementation of the UNDP’s Cash for work. The UNDP implementation of the post Matthew cash for work has involved local authorities across the different activities of the program, building long-term local capacity. In particular, UNDP have allocated to the local authorities of each intervened municipality in between 7 and 8% of the total amount of the cash for work resources to be spent in the municipality. The variation between 7 and 8% is due to the total cost of the implemented modality of payment. As such, 40% of the total cost of the program in the municipality has always been allocated to job creation. The amount to be spent by the local authorities, have been useful for them to meet their needs. Interviews with Mayors and advisor of the benefiting municipalities pointed out that local authorities feelt empowered with this transfer of resources directly to them. Given the very small tax revenues that local authorities collect and receive, the cash for work amount help local authorities to exercise their own identification and prioritization of needs and also improved their decision making capacity to spend this amount according to their needs. For instance, in the case of Jeremie, the major, stated this amount was used to do extra-cash for work required and operated by the municipality. Similarly, the Dan-Marie major stated the amount helped to contract further personnel to supervise works and ensure the final community assets where cash for worked was used resulted with even better standards than the ones held before the hurricane. Still, as monitoring and controlling strategy to avoid misuse of this amount, the program requires local authorities to report back the purpose of the expenditure. 3.9 Allocation of resources and geographical coverage of the program The UNDP program was conceived as a flat benefiting amount per municipality. In average each municipality got allocated an amount of 50 thousand USD. For operative purposes, this amount was fixed disregarding the size of the municipality or the impact of the hurricane. 26 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) Tailoring this amount to the size of the municipality and focusing in particular section communales where the impact of the hurricane was the greatest could conduce to maximize the impact of the program. In this case, the available information that can be used for this purpose is the 2015 forecast of population from the National Institute of statistics, which offers information at the section communal level. Also, the poverty figures that the 2012 ECVMAS provide by department and rural and urban areas. And finally, the Rapid post-disaster assessment of needs. In fact, this post-disaster assessment has helped the local authorities to identify the sites to intervene. So, no unfruitful infrastructure intervention was done in the case of the UNDP cash for work. However, during the cash steering committee meetings some partners shared experiences where unfruitful interventions in some cases where found. 4 Other type of cash interventions: A multipurpose grant size design for Haiti This section presents a proposed design for the size of a multipurpose cash grant done jointly between OCHA and PNUD in the context of the cash steering committee in the post Mathew activities. The objective of this proposed multipurpose grant is to contribute towards the meeting of the basic needs of the most deprived and vulnerable, crisis-affected population in the Sud and Grand’Anse departments. The design the size of this multi purpose grant was based on information available in the country with regards to the minimum expenditure basket. The proposed value takes into account departmental differences when possible and the alternatives available in the country to update this value because the fluctuations of the prices. The proposed transfer value aims to account for the food and non-food items that compose a minimum expenditure basket here in Haiti. It uses the analysis done by the World Bank (WB) and the UOPES, when developed the extreme and moderate poverty lines for Haiti. To develop both poverty lines, the 2012 ECVMAS was used. This survey comprises information representative by department on household’s income and consumption. The WB and UOPES analysis use such information to develop a national minimum expenditure basket. 27 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) On the other hand, Haiti also has available for public use the food basket developed by the CNSA. Given that this food basket is updated monthly gathering information nationwide, the proposed size of the multipurpose grant departs from this basket so the basket can be updated easily every month since now. This CNSA food basket at November 2016 takes a value of 1,160HGD per person per month. It consists of six products: Rice, Wheat flour, Maize, Beans, Sugar and Vegetable oil, which all together give a caloric value of 1979 calories per person per day. In use of this basket, prices can be monitored by the existing monitoring activities and could take into account possible abrupt fluctuations given in times of crisis. However, this basket only comprises the food items. Then, to take into account the non-food items in the proposed multipurpose grant, we use the 2012 ECVMAS information to determine the percentage of expenditure that an average household in Haiti spends in food. Figure 7 below presents this percentage by department. Figure 7. Engel coefficient: Percentage of food expenditure in the total household expenditure by department Source: ECVMAS, 2012 To avoid determining this percentage for the whole distribution but rather for a group of population that represents the medium household, we calculated this Engel coefficient as the mean across households in between the 20 and the 60 percentile and focusing exclusively on households in the Grand’Anse and Sud departments. The obtained percentage corresponds to 67.8%. 28 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) We applied this estimated Engel coefficient over the value of the CNSA food basket to determine the cost of the non-food share of the minimum expenditure basket. As such, our estimated value for the minimum expenditure basket is 1,712 HGD per person per month. Still, from this obtained value we discounted the average percentage value paid in rent by this group of population (4.6%) and also the average percentage value that could come to the household by transfers from other households (18.2%). As such, we obtained as total per capita average size grant 1,322HGD per person per month. Table 4 below summarizes all these results. Table 4. Multipurpose grant size design, HGD per month per person Source: author’s calculations based on CNSA and 2012 ECVMAS However, we propose to take into account the different sizes of the households to avoid inefficiencies in the grant size. According to the 2012 ECVMAS 45% of the population in the Grand’Anse and Sud department are living in households made of 4 to six persons per households; 32% in households made of 1 to 3 persons; and 23% of the population living in households consisting on 7 or more persons. Then, the proposed multipurpose grant size is 40USD per month per household made in between 1 and 3 persons per household, 100USD per month per household consisting on 4 to 6 persons; and 160USD per month per household consisting on seven or more persons (Table 5). 29 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant)  Table 5. Proposed multi-purpose grant value per month per household Source: Author’s calculations based on ECVMAS 2012 and CNSA. As observed in this design we adjust the grant size across time and geographical area, so it reflects the fluctuation of the prices in the local markets. We also tailored different grant sizes according to different types of households/potential beneficiaries. The proposed guidelines to use this grant size are the following: • Monitor monthly variation in prices in the food basket using the CNSA study. • As CNSA tracks only the food basket prices, assume the prices of the non- food items vary at the same rate as the food items. • Keep the multipurpose grant transfer value fix and update the value every six months according to the cumulative variation in the food basket prices for the past six months. • However, if the food basket price increases more than 5%, update the value of the grant as per the observed variation. This means that we are proposing here to set as cut-off point limits for maximum acceptable price inflation, 5%. Finally, in terms of the number of instalments to be given via the multipurpose grant, we left this open to be defined by the possible program according to the availability of resources. 30 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant)  5 Remarks and steps ahead This document presented and overview of the available international evidence on cash transfer programing, the lessons learnt from the first six months of the post Matthew Haiti’s UNDP cash for work. It used quantitative and qualitative assessments to draw conclusions on these lessons. From the international lessons we learns that Cash transfer interventions have became during the last years one of the most preferred modalities because their proven ability to meet particular program objectives, while giving flexibility to encompass beneficiaries’ preferences. But, no golden rules to follow exist, rather international cash lessons have been learnt. In particular, documented impact evaluations were found on the analysis of sectors such as Food security and nutrition, Livelihoods, Shelter, Distribution of non-food items, Access to basic services, Household decision-making, Gender equity and empowerment. This international evidence have shown us that cash transfers are relevant whenever beneficiaries of the cash intervention are able to buy what they need in the local markets, at reasonable prices. Also, cash must be able to be delivered and spend safely. In terms of the experience worth sharing as lessons learnt from the cash programing in post Matthew Haiti. First, targeting has proved to be a fundamental piece that could boost the humanitarian coordination of cash actions, allowing maximizing the impacts on the most needy beneficiaries. However, post Matthew Haiti failed into achieving a coordinated targeting of potential beneficiaries and not even sharing of list of beneficiaries was possible. This topic remains on the agenda for future preparedness actions for emergency contexts. The Matthew emergency has highlighted its relevance and the possibilities that efforts in this direction could bring. Second, despite the different humanitarian actors have agreed some standards in terms of the amount to be paid as wage in cash for work programs, yet variations in the daily wage given to beneficiaries are found. But, there is yet no evidence in Haiti of disruption of labor markets because of the cash for work activities. The formal minimum wage taking into account social security is still higher than the agreed value to pay per day to beneficiaries. Nonetheless, this minimum wage is dissonant and significantly lower than the value of a subsistence food basket. As such, one recommendation to fill in this gap and assure programmatic impacts is to implement slot sizes per worker that would allow covering the household minimum food basket. Third, gender perspectives as the implemented by the UNDP cash for work program and even further actions could prove to have important impacts on 31 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) women empowerment and rupture of gender work stereotypes. Different qualitative assessments of the UNDP gender deferential activities suggest promising impacts. Fourth, further work is required to evaluate the cost/benefit of using different financial partners to operationalize the cash payments. This analysis must take into account not only the financial cost but also the indirect economic costs that each modality of payments could bring, such as security breaches. This report also analyzed and proposed a grant size for future multipurpose grants in the country. Further work is required in this line of action to determine in which contexts this type of intervention could be more effective to be applied. Still, recommendations were derived to update and operationalized this grant size over the Haitian territory. This report is part of a group of reports to be done by the different implementing agencies to evaluate the first six months of humanitarian assistance in cash transfers. It is meant to complement these other reviews of lessons learnt from cash transfer programs. References ActionAid (2008). Cash programming: An experience from the Andaman and Nicobar islands. Bailey and Walsh (2007). The use of cash transfers in Emergency and Post- Emergency Non-Food Item Programs, Journal of Humanitarian Assistance, http://jha.ac. Devereux et al. (2006). After the FACT: An evaluation of Concern Worlwide’s Food and Cash Transfer Project in Three Districts of Malawi, A Report for Concern Worldwide by the Institute of Development Studies. Dunn (2009). External evaluations: Fresh food voucher project by Action Against Hunger, Dadaab Refugee Camps, Kenya. Frize (2002). Review of the cash for work component of the drought recovery programme in Turkana and Wajir Districts (September 2001 – June 2002). Harvey and Baily (2011). Cash transfer programming in emergencies. Commissioned and published by the Humanitarian Network at ODI. Harvey and Morongwe (2006a). Independent evaluation of Oxfam GB Zambia’s emergency cash transfer programme. ODI for Oxfam. 32 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant) Harvey and Morongwe (2006b). Independent evaluation of Oxfam GB Zambia’s emergency cash transfer programme. University of Arizona, Phase II Monitoring and evaluation Project for the Tanout Cash Distribution Project. Jaspars, Harvey, Hudspeth and Rumble (2007). A review of UNICEF’s Role in cash transfers to emergency-affected populations, EMOPS Working paper. Kardagan et al. (2010). Evaluation of the Zimbabwe’s Emergency Cash Transfer (ZEST) Programme. Concern Worldwide and Oxford Policy Management. OCHA (2017). Hurricane Matthew humanitarian dashboard as of February 2017. Sandstorm and Tchatchua (2010). Do cash transfers improve food security in emergencies? Evidence from Sri Lanka. In Omamo et al. (eds), Revolution: From food aid to food assistance. Rome, WFP. Skoufias, Tiwari and Zaman (2011). Can we rely on cash transfers to protect dietary diversity during food crisis? Estimates from Indonesia. Washington DC, World Bank. WFP (2009). Cash and Vouchers Manual. First Edition, Rome. Annexes Table A1. CaLP possible contacts Nathalie CISSOKHO Thérèse Faye CaLP West Africa Focal Point CaLP West Africa Regional Administration Mobile phone: (+ 221) 77 333 49 56 and Communication Officer ACF: Rue 50 Yoff Virage – Mobile phone: (+ 221) 77 819 49 04 Dakar – Senegal ACF: Almadies, lot B, Rond Point Ngor – B.P. 29 621 ; Dakar Yoff Dakar – Senegal Email: westafrica@cashlearning.org B.P. 29 621 ; Dakar Yoff Skype: nathalieklein2 Email: admin.westafrica@cashlearning.org www.cashlearning.org Skype: theresediouma www.cashlearning.org 33 Early lessons learnt from cash interventions in post Matthew Haiti Yadira Díaz (Consultant)